Reversing the hold up vs hold out debate?
Exactly 7 years ago, on 16 July 2015, the CJEU rendered its Judgment in Huawei v ZTE (here are the comments I published that day).
The Huawei v ZTE Judgment essentially sought to clarify the circumstances under which the seeking of injunctions by a SEP holder could constitute an abuse of dominance. The Judgment confirmed the view, initially advanced in academic circles, and endorsed by the European Commission in Samsung and Motorola, (and vehemently opposed by many) that in certain cases patent hold up was a competition law problem connected to the leveraging of market power obtained through standardization. The underlying idea was that hold up could materialize in refusals to licence, excessive royalties or injunctions. In that Judgment the Court set up a procedural framework balancing the different stakes and incentives at issue.
7 years later many of these debates remain (and remain equally bitter). Interestingly, though, there appears to have been an effort to shift attention away from hold up and focus, instead, on hold out (i.e. the situation where implementers would allegedly refuse to negotiate in good faith). The argument is that innovation on the part of SEP holders would be discouraged should their royalties not be high enough as a result of hold out.
Paying attention to potential hold out on case-by-case assessments might be, to some extent, natural because implementing the procedural framework set out in Huawei v ZTE necessarily requires assessing whether implementers have entered into bona fides negotiations.
At the same time, however, the recent trend is to present hold out (aka “reverse hold up”) as the other side of the same coin. This view has made it from economic articles, to national litigation, to the “new Madison” policy in the US under AAG Delrahim. More recently, and more surprisingly, the European Commission’s draft horizontal Guidelines (recital 470) would appear to support this view:
“When the standard constitutes a barrier to entry, the undertaking could thereby control the product or service market to which the standard relates. This in turn could allow undertakings to behave in anti-competitive ways, for example by refusing to license the necessary IPR or by extracting excess rents by way of discriminatory or excessive royalty fees thereby preventing effective access to the standard (“hold-up”). The reverse situation may also arise if licensing negotiations are drawn out for reasons attributable solely to the user of the standard. This could include for example a refusal to pay a FRAND royalty fee or using dilatory strategies (“hold-out”)”.
Perhaps it is simply a drafting problem, but this paragraph appears to put hold up on the part of SEP holders and hold out on the part of individual users of the standard (and the concerns to which they both relate) at the same level, also from a legal standpoint. This is interesting for various reasons that we have often discussed on this blog. First, the shift in the focus of these debates is one more example of the pendulum oscillations that characterize competition law, but one where the swing would appear to be particularly wide. Second, this text would also appear to equate hold out practices with anticompetitive hold up practices on the grounds that both can affect the distribution of rents between the different parties, regardless of whether they involve the exercise of market power or not.
I would welcome your views on this point. Not having worked for clients on these issues, I have no view on the extent to which hold out may be a real-life concern. As a competition lawyer, however, I have trouble seeing how hold out practices could lead to genuine competition law concerns (i.e. how they could lead to foreclosure, anticompetitive leveraging, exploitation or otherwise restrict competition) absent dominance or a cartel/boycott-like arrangement at the level of would-be licensees. I see that others have expressed very similar thoughts (e.g. here or here).
Don’t get me wrong. As mentioned above, hold out considerations can be, and have been, relevant in case-by-case assessments under the Huawei v ZTE framework (under that framework injunctions remain legitimate in relation to implementers not acting in good faith). But to the extent that hold out concerns may be concerned with relative bargaining power (as opposed to market power) and with the distribution of rents between SEP holders and implementers (absent market power, exploitation or foreclosure), they would not appear to be a matter for competition law to address. In sum, while the narrative, the incentives, and perhaps even the economics, may be “the reverse” as those arising in hold up scenarios, this might not be accurate from a legal standpoint.
Chillin'Competition 
[second link doesn’t work?]
I have been involved with this since the early 1990s, and still am. I have seen both holdup and holdout over the years, albeit holdup quite a bit more often than holdout. Here is an excellent perspective by Carl Shapiro and Mark Lemley — both excellent commentators with deep experience in this area. https://faculty.haas.berkeley.edu/shapiro/patentholdup.pdf.
Some salient quotes, where they put their finger on the sore spot: “Patent advocates have sought to deflect concerns about patent holdup not only by denying its existence but by concocting a supposedly parallel story of “patent holdout.” On this theory, patent owners are being deprived of the fruits of their R&D investments by implementers who copy their technology but refuse to pay. The idea is to tell a story that parallels patent holdup”. Keep reading on p. 2049 — although the whole article is excellent. Then they add: “Those who express concerns about patent holdout seem to want to increase the returns to patent holders whose inventions add little or no incremental value. That’s simply not how the patent system works or is intended to work. Indeed, doing so would create perverse incentives for companies to seek patents with holdup power rather than to fund R&D programs leading to technological advances.” And indeed, I have seen this in standard setting: companies will include other’s patents in a standard in exchange for the other agreeing to put their technology in as well (or in another standard). This patent stuffing and horse trading (or “I’ll scratch your back if you scratch mine”) can be a real problem, and can turn standards into royalty capture devices, that have little or nothing to do with innovation incentives, and in some cases add zero value to consumers.
Conclusion: my experience tallies with yours, that holdout does not raise competition concerns, and can be effectively addressed by following the “dance” described in Huawei v ZTE, and by “FRAND injunctions” along the lines of what the UK Supreme Court blessed in Unwired Planet v Huawei. (But I do have some points of criticism of this judgment, here: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3872755).
Maurits Dolmans
19 July 2022 at 3:13 pm
Dear Alfonso,
The questions you raise are important and the issue you raise concerning §470 of the draft horizontal guidelines does not seem to be a mere drafting problem. I think it originates in the discussion on what the goal of (EU) competition policy is: is it to protect the competitive process as such or is it to protect consumers, i.e. to protect competition in the interest of the customers of the products concerned.
For those who believe the goal is to protect the competitive process as such, collective purchasing and buying power are equally bad as collective selling and market power on the sales side (sometimes referred to as the “mirror theory”). The competitive process is easily harmed, simply because as a result of collective purchasing there are less (independent) buyers on the upstream market.
For those, like me, who believe that the goal is and should be to protect consumers, collective purchasing and buying power are generally only a possible competition problem if the collective purchasing or buyer power leads to a reduction of competition downstream and/or negative effects for the customers downstream. Here the theories of harm are much less self evident. One could think of buying power being used to lower the input price for some but leading to a higher price for others (‘waterbed effect’) and thereby limiting competition downstream. Or collective purchasing leading to such a price squeeze upstream that it does not only lead to a different distribution of rents away from the upstream suppliers but would also undermine their innovativeness, with the latter effect resulting in harm for downstream customers. However, I don’t recall these theories of harm ever being shown in actual cases.
In terms of ‘hold up’ versus ‘hold out’: ‘hold up’ is linked to theories of harm of the refusal to supply and excessive pricing categories while ‘hold out’ can, in theory at least, be linked to refusal to buy and excessively low pricing theories of harm. While the former theories of harm are applied in a non-insignificant number of important cases, the latter theories of harm have, as far as I know, never been used in actual cases.
I guess what you found in §470 is a lingering on of that old debate on the goal of EU competition policy, a debate that should have ended by now but that nonetheless may have left this and other traces and remnants in the draft horizontal guidelines, in particular in the chapter on collective purchasing.
Luc Peeperkorn
19 July 2022 at 4:41 pm
I disagree with the suggestion that while hold-up causes consumers harm in the form of higher prices, hold-out does not harm consumers. Competition law rightly condemns coordinated price fixing both by sellers (resulting in uncompetitively high prices) and by buyers (resulting in uncompetitively low prices). That is because both ultimately harm the competitive process and therefore consumers. Likewise, there is no principled basis for distinguishing the harm to consumers caused by uncompetitively high licensing rates and the harm caused by uncompetitively low licensing rates. Consumers can be harmed by artificially reduced innovation and choice at least as much as they can be harmed by higher prices.
Fabian Gonell
20 July 2022 at 4:56 pm
Dear Fabian, thanks for your reaction.
From your reaction to me and your other reaction to Alfonso, I see you are a firm believer in the (outdated) ‘mirror theory’: that competition law should equally condemn price fixing by sellers and buyers, as both harm the competitive process “and therefore consumers”.
I see how you can get to your position by taking protection of the competitive process ‘as such’ as the goal of EU competition policy, leaving it open which benchmark to use deciding whether the competitive process is harmed, and considering it sufficient to simply assume (instead of show) that consumers are harmed.
However, if it is considered, as confirmed by recent case law, that the goal of EU competition policy is to protect competition for the benefit of consumers, that is to avoid negative effects for the customers downstream, then there are good reasons to differentiate the treatment of price fixing by sellers and price fixing by buyers. Price fixing by suppliers will increase the sales price and thereby directly lead to harm for the customers downstream, while countervailing efficiencies are generally unlikely. Cooperation by buyers to lower the purchase price of an input does not lead to a direct harm for the customers downstream; it may even lead to a benefit for these customers, to the extent that competition between the buyers forces them to pass on (part of) the lower input price and/or if it allows these buyers to compete more forcefully with other, possibly bigger, buyers who already obtain the lower input price. This does not mean that cooperation between buyers can never have anti-competitive effects for the downstream customers, including, as indicated already in my reaction on Alfonso’s piece, that collective purchasing may not only lead to a different distribution of rents away from the upstream suppliers, but could also undermine their innovativeness, with the latter effect resulting in harm for downstream customers. However, such negative effect on innovation cannot simply be assumed, as you seem to do, but will need to be substantiated and shown to at least balance the expected positive effects resulting from a lower input price. Using your words, yes there is a “principled basis for distinguishing the harm to consumers caused by uncompetitively high licensing rates and the harm caused by uncompetitively low licensing rates.”
Luc Peeperkorn
20 July 2022 at 7:38 pm
To borrow a term made famous (at least in our small world) by Professor Philip Areeda in an article on Essential Facilities (here: https://www.jstor.org/stable/40843140), “hold-out” is truly “an epithet in need of limiting principles”. What we see in the trenches of SEP policy discussions is that SEP licensors like to use the term “hold up” to describe a wide range of licensee conduct, some of which may be dilatory and unreasonable, but much of which is ordinary-course diligence that prudent commercial parties engage in before agreeing to make what may be significant payments (individually or in the aggregate) to SEP licensors. For example, a prospective licensee may take the time it needs to do detailed analysis of whether some or all of the patents it is offered for license are valid and infringed by the practice of a standard. A putative licensee that does business globally may want to evaluate questions of validity and infringement under different national patent laws in various countries where sells products that implement the standard as to which the putative licensor claims to own valid and essential patents.
All of this takes time, and, like any seller, licensors prefer to get paid sooner rather than later. So they try to describe even ordinary-course licensee diligence as “hold-up”, perhaps in the hope that they can persuade courts in jurisdictions where injunctions in patent cases are granted more readily that a putative licensee is engaging in delay and should be enjoined from selling products that implement the standard. That threat may cause putative licensees to cut short their ordinary-course diligence, which is the point at which the “hold-up” epithet is used to chill the kind of behaviour that, in any other commercial context, competition law views as pro-competitive.
Competition enforcement agencies have a role to play in identifying kinds of licensee diligence that fall within a safe harbour and, therefore, can never be the basis for an injunction aimed at coercing a putative licensee to take a license on super-competitive terms. The draft policy statement released late last year by the Antitrust Division, the Patent and Trademark Office, and NIST in the US seemed to recognize that this was a helpful role for competition enforcers to take, but the withdrawal of the draft leaves a void in this respect, one that enforcers in the US, Europe, and elsewhere could usefully fill. After all, would-be licensees that do the hard work of invalidating patents that should not have been granted are clearing a path for innovation, and should be encouraged, not coerced into abandoning their good work through the threat of injunctions.
Gil Ohana
19 July 2022 at 8:26 pm
Oops, corrected a small but significant typo in my original post.
********************
To borrow a term made famous (at least in our small world) by Professor Philip Areeda in an article on Essential Facilities (here: https://www.jstor.org/stable/40843140), “hold-out” is truly “an epithet in need of limiting principles”. What we see in the trenches of SEP policy discussions is that SEP licensors like to use the term “hold out” to describe a wide range of licensee conduct, some of which may be dilatory and unreasonable, but much of which is ordinary-course diligence that prudent commercial parties engage in before agreeing to make what may be significant payments (individually or in the aggregate) to SEP licensors. For example, a prospective licensee may take the time it needs to do detailed analysis of whether some or all of the patents it is offered for license are valid and infringed by the practice of a standard. A putative licensee that does business globally may want to evaluate questions of validity and infringement under different national patent laws in various countries where sells products that implement the standard as to which the putative licensor claims to own valid and essential patents.
All of this takes time, and, like any seller, licensors prefer to get paid sooner rather than later. So they try to describe even ordinary-course licensee diligence as “hold-up”, perhaps in the hope that they can persuade courts in jurisdictions where injunctions in patent cases are granted more readily that a putative licensee is engaging in delay and should be enjoined from selling products that implement the standard. That threat may cause putative licensees to cut short their ordinary-course diligence, which is the point at which the “hold-up” epithet is used to chill the kind of behaviour that, in any other commercial context, competition law views as pro-competitive.
Competition enforcement agencies have a role to play in identifying kinds of licensee diligence that fall within a safe harbour and, therefore, can never be the basis for an injunction aimed at coercing a putative licensee to take a license on super-competitive terms. The draft policy statement released late last year by the Antitrust Division, the Patent and Trademark Office, and NIST in the US seemed to recognize that this was a helpful role for competition enforcers to take, but the withdrawal of the draft leaves a void in this respect, one that enforcers in the US, Europe, and elsewhere could usefully fill. After all, would-be licensees that do the hard work of invalidating patents that should not have been granted are clearing a path for innovation, and should be encouraged, not coerced into abandoning their good work through the threat of injunctions.
Gil Ohana
19 July 2022 at 10:50 pm
Very interesting post, and I appreciate your call for others’ views. Having been “the client,” or at least one of lawyers working at the client (Qualcomm), I offer the following thoughts:
Hold out’s greater real-world impact justifies its increased attention
The increasing attention to hold-out is most easily explained by the fact that hold-up is a largely solved problem, while hold-out, which has been increasingly evident over the past decade, persists.
With respect to hold-up, whatever may have been true 10 or 15 years ago, it is now clear in every major jurisdiction that injunctions are not available against an implementer that is a willing licensee – that is, a licensee who is willing “must be one willing to take a FRAND licence on whatever terms are in fact FRAND”. (Unwired Planet) That is certainly true in the US, where even the ITC considers whether an infringer is a willing licensee in determining whether an exclusion order is appropriate. Thus, in every major jurisdiction there is basically ZERO possibility that a patent holder can use the threat of exclusion to coerce unfairly high royalties. The infringer merely need file a claim or counterclaim for breach of the FRAND commitment, and seek specific performance of that commitment. If the infringer is wrong, and the patent holder has made a FRAND offer, then if the infringer is a willing licensee it will accept that offer. If the infringer is right, then the patent holder will not be able to exclude the infringer until it makes a FRAND-compliant offer. In either case, there is no mechanism by which the patent holder can effectively coerce the infringing implementer. In other words, to the extent hold-up was ever a real problem in SEP licensing, the only mechanism to impose that strategy was exclusion, and courts around the world have adopted mechanisms, centered around the concept of willing licensee, that have addressed it.
In contrast, it has been much more difficult for Courts to adopt simple rules that address hold-out, because hold-out largely involves unilateral gamesmanship by infringers, such as delaying license negotiations or trying to insist on patent-by-patent, jurisdiction-by-jurisdiction litigation of large SEP portfolios to delay and drive up costs for inventors, rather than the need for an action by a neutral adjudicator as is the case with an exclusion order. To be sure, Courts have made some progress, mainly by defining certain hold-out actions as manifesting unwillingness to take a FRAND license. But bad-faith implementers continue to resist these measures, arguing instead that their mere declaration of willingness to take a license to any patent – but only those patents –that have been fully litigated through all appeals qualifies as “willingness”. What this would mean in practice is that SEP holders would have to litigate every single patent in their portfolio at the request of the implementer – no portfolio agreement – even if it takes 20 years, with no possibility to argue to a court that the implementer’s recalcitrance is causing irreparable harm warranting an injunction.
As a result of this imbalance, hold-out has a much greater real world impact than hold-up. In the past several years, there has not to my knowledge been any significant implementer threatened with an injunction or exclusion order for infringement of a cellular SEP that has not had the opportunity to avoid that injunction by demonstrating that it is a willing licensee. In contrast, several cases have found licensees “unwilling” and the dispute between Apple and Qualcomm involved overwhelming hold-out pressure.
If hold up implicates competition laws, then so must hold out
Whether hold up or hold out is a problem is a different question than whether they are competition problems. From a competition law perspective, the suggestion that hold-up by patent holders raises competition or antitrust concerns, while simultaneously hold-out by implementers does not, cannot be true. If hold-up implicates the competition laws, the hold-out must also necessarily implicate those laws. If hold-out does not implicate the competition laws, then hold-up must not either.
Let me explain why. For purposes of this discussion, assume that in either the case of hold-up or hold-out, nothing wrongful has happened during the standardization process — both implementers and innovators acted in good faith during standardization, and no deception was involved. So you have collective action at time “A” – standardization – that was lawful and pro-competitive. Later – maybe years later – you have a unilateral action at time “B” that is either hold-up or hold-out.
Focusing on “hold-up” first, there can be no serious suggestion that hold-up (using the threat of exclusion to obtain unreasonably high royalties) is a violation of the Sherman Act. Courts in the US have specifically held that excessive pricing by itself does not violate the Sherman Act, and going to court or the ITC to seek an injunction cannot violate the Sherman Act under the Noerr-Pennington doctrine unless the case is objectively baseless. Nor is there precedent in EU competition law that claims of pure excessive pricing in relation to SEPs constitutes a competition law offense. And we should be past the point that passing references to theoretical “foreclosure, anticompetitive leveraging, [and] exploitation” suffice to support the idea that “hold up” – without more – harms the competitive process. Instead, the suggestion is that, as one regulator put it, “SDO members engaging in concerted activity and choosing between substitute technologies,” — that is, standardization itself before any hold-up activity by a SEP owner – renders the post-standardization act of hold-up subject to competition law scrutiny.
With that understanding of the purported basis for competition concerns about hold-up, let’s turn to hold-out. I’ve seen statements that hold-out is not an antitrust concern because it is unilateral conduct.
Simply saying that hold out is unilateral, however, does not distinguish it from hold up. Hold up is also a unilateral act. So if hold out is not an antitrust concern because it is “unilateral”, then hold up isn’t an antitrust concern either, because it is also unilateral. In both cases there are potential remedies under patent and contract law.
If, on the other hand, hold up is “an abuse of market power to stymie consumer choice,” as has been suggested, then the same necessarily applies to hold out. In cases where, as a result of competition, a specific standard occupies the field, then market power exists in the market for technologies in the field occupied by that standard before future versions of that standard are finalized. Abuse of the FRAND commitment to hold out is abuse of that market power (and here I explicitly disagree with your reference to “relative bargaining power”, suggesting that it is immune from competition law query). And that abuse unquestionably stymies innovation and consumer choice, because it has a profound impact on the incentives of companies competing in the innovation markets that consumers depend on for technological advances. It is as much an error to say that hold-up “raises competition concerns” and hold-out does not as it is to suggest that price fixing by sellers is unlawful but price fixing by buyers is not.
To be sure, whether either hold-up or hold-out is actually appropriately addressed by competition law remains an open question. There are contractual and patent law remedies available that fully address these issues, and creating potential antitrust liability for a unilateral act because it has some relationship to an otherwise lawful prior collective action years earlier could introduce a great deal of uncertainty and instability into both antitrust law and the marketplace. But there is no basis in competition law for treating hold up and hold out differently.
Fabian Gonell
20 July 2022 at 4:52 pm